Manipulations of fundamental frequency and formant dispersion influence
attractiveness of male voices: Female preferences for testosterone
dependent traits in male voices. 

Feinberg, DR., Jones, BC., Little, AC., Burt, MD., Perrett, DI. 

Perception Laboratory, School of Psychology, University of St Andrews

email: drf3@st-and.ac.uk

We consider the possibility that the physical structures of the vocal
apparatus allowing human speech may have evolved via the processes of
sexual selection.  Recent research has shown that female preferences for
males with low fundamental frequencies may be potentially adaptive. 
Here we manipulated 2 sexually dimorphic properties of real male voices
(fundamental frequency and formant dispersion) separately and in
conjunction to isolate the effects of source (fundamental frequency) and
filter (formant dispersion) characteristics on vocal attractiveness,
perceived vocal health, masculinity, size, and age.  Females rated male
voices with lowered fundamental frequencies as more attractive,
healthier, more masculine, larger, and older than rated male voices with
raised fundamental frequencies.  Male voices with decreased formant
dispersion were rated as more masculine, larger and older, but not more
attractive or healthier than male voices with increased formant
dispersion.  Only larger females preferred the male voices with
decreased formant dispersion, suggesting assortative preferences for
male size as indicated by filter characteristics of the vocal tract.  A
combined manipulation of raised fundamental frequencies and increased
formant dispersion decreased female ratings of attractiveness, health,
masculinity, size, and age.  In normal males, pubertal testosterone
levels determine fundamental frequency.  Immunocompetence theory states
that only the fittest males should be able to afford the
immunosuppressive effects of testosterone.  Therefore, males with
relatively high testosterone should be most attractive because they
should be healthy, dominant individuals.  Here, we show support for this
theory because simulating the effect that testosterone has on the male
voice increased perceptions of attractiveness and health, whilst
simulating an increase in vocal-tract length had no overall affect on
perceptions of attractiveness and health.